Version, test, and compare prompts across GPT-4o and Claude side-by-side. Track costs per run. Find what actually works.
system: You are a helpful assistant. specialized in {domain}. user: Summarize the following in {tone} tone: {content} ▶ GPT-4o · 142 tokens · $0.0014 · 890ms
features
Variables with {curly_braces} — define inputs once, test across every variant automatically.
Run the same prompt against GPT-4o and Claude simultaneously. See outputs diff'd side-by-side.
Per-run token counts and USD cost. Know exactly what each prompt iteration is costing you.
Every edit is stored. Roll back to any version, compare diffs, and audit what changed.
pricing
self-hosted
forever. MIT license.
cloud
no ops, just prompts.